This invention relates in general to particle analysis. More particularly, it relates to methods and devices for performing automated blood cell analysis by integrating "impedance," "light scattering," and "fluorescence" analysis and flow cytometric techniques. This invention also relates to a multipurpose reagent system and a method for rapid analysis of a whole blood sample.
Peripheral blood of a human usually contains red blood cells (RBC), platelets (PLT), and white blood cells (WBC), all of which are suspended in a conductive medium commonly known as plasma. Plasma comprises proteins, anions and cations. Plasma also contains components which assist in forming blood clots.
The blood in an adult usually contains about 4.5 to 5 million RBCs or erythrocytes per cubic millimeter. Mature RBCs have no nuclei and are generally shaped as circular biconcave disks with a diameter of about 7.5 to 8 microns (.mu.), and a thickness of about 1.5 to 1.8 microns. RBCs contain hemoglobin which gives blood its red color. Hemoglobin helps transport oxygen and carbon dioxide and plays a role in maintaining pH in blood.
The blood in an adult usually contains about 200,000 to 400,000 platelets per cubic millimeter. Platelets are small, biconvex cellular particles whose mean volume is about 7.mu. to 8.mu.. Their general configuration includes a granular central portion embedded in a homogeneous matrix.
Peripheral blood also contains red cells of earlier maturation levels which are important diagnostic indicators. Two of these are reticulocytes and nucleated red blood cells.
At the earliest stage of development the red cell consists mostly of nucleus, and is referred to as an erythroblast. As the erythroblast matures, the nucleus becomes smaller, anucleolate, and more nearly spherical. Subsequent maturity involves a complete loss of nucleus. The immature red cells that retain a nucleus are referred to as nucleated red blood cells (NRBCs). The NRBC count has been useful in patient monitoring under many disease states. However, NRBCs in peripheral blood often contribute to inaccurate enumeration of the white cell count, due in part to the presence of a nucleus which makes them difficult to distinguish from small white cells.
Reticulocytes are red cells at the maturation level just between NRBCs and mature RBCs. Reticulocytes provide a means of evaluating a patient's anemic state. Anemia usually occurs as a result of an uncompensated increase in the rate of removal of erythrocytes from blood, or a decrease in the rate at which they are formed and released into blood. An increased reticulocyte patient count in an anemic patient indicates rapid erythroid turnover which suggests acute blood loss or hemolysis.
In normal human blood, the concentration of white cells, referred to as WBCs or leukocytes, is much lower than the concentration of red cells. The normal concentration of WBCs is approximately 7000 per microliter. They vary in size, most of them from about 7.5 to 12.5 microns in diameter. They are more nearly spherical in shape than RBCs and usually somewhat larger in volume. WBCs may be classified generally as either granular or non-granular. The granular WBCs include neutrophils, eosinophils and basophils. The non-granular WBCs include monocytes and lymphocytes. These categories of WBCs are often referred to collectively as a "five-part differential," and, generally, the most significant of these categories are neutrophils and lymphocytes.
Neutrophils usually comprise from about 50 to 60% of all WBCs. Their cytoplasm contains numerous minute granules which can be stained. Under certain conditions neutrophils may leave the blood vessels and disintegrate, thereby releasing granules into the connective tissues. These granules are rich in certain enzymes which become active and take part in the body's defense mechanism.
Lymphocytes comprise about 30% of the WBCs in humans. The nucleus of a normal lymphocyte occupies nearly the entire cell volume, and thus the cytoplasm surrounding the nucleus is a rather thin shell. Lymphocyte cytoplasm may stain with dyes due to the cytoplasm's content of ribonucleic acid.
Lymphocytes may leave the blood vessels and enter the connective tissue where they also constitute a part of the body's defense mechanism, playing a major role in the body's immunological responses.
There are three major "subsets" of lymphocytes that are currently clinically significant: T lymphocytes, B lymphocytes, and Natural Killer cells, also known as "large granular lymphocytes" or NK cells. Each of these subsets can be distinguished based on the existence of distinctive cell surface markers or antigens. Also, B lymphocytes have a high density of immunoglobulin of their surfaces, whereas T lymphocytes have little or none. T lymphocytes are characterized by various surface markers against which antibodies can be produced.
Categories of T lymphocytes have been identified according to their surface markers and overall function. The "helper" T cells help B cells produce certain classes of antibody molecules, and help other T cells in their immune responses. The "suppressor" T cells are regulatory cells that can suppress the responsiveness of other T or B cells. The suppressor T cells include several subsets which are also recognized by distinct surface markers.
The ability to count, size and classify blood cells is useful when evaluating the health of an individual. For example, the level of circulating CD4 lymphocytes (helper-T cells having a CD4 antigen expressed on the surface of the cell) is currently regarded as the best single predictor of progression of HIV infections. The CD4 level may be used for classifying individuals for enrollment in experimental treatment regimes, determining when anti viral therapy should be initiated, and monitoring treatment responses in clinical trials. Because CD4 lymphocyte levels may be important to some HIV-infected individuals, it is desirable to measure this parameter accurately.
In the current state of the art of cell analysis, there are two technologies used for counting and classifying cells. These are generally known as "flow cytometry" and "image cytometry." The flow cytometry technology, which essentially consists of passing cells one at a time through a sensing zone of a flow cell, is preferred in clinical applications where patient test load is an important metric. This is mainly because it has at least an order of magnitude advantage in the number of cells that can be analyzed per second.
Instrumentation incorporating flow cytometry can be further subdivided into two methods which can be generally classified as "conventional hematology" and "fluorescence cytometry."
A primary distinction between the two methods is that conventional hematology generally distinguish cells by means of size and shape alone using primarily impedance and light scatter technologies, whereas fluorescence cytometry uses cell nucleic acid content and/or surface antigens in addition to size and shape in distinguishing cells. Therefore the fluorescence method may be used to subdivide the cell types into finer classifications.
A second distinction between the two methods is that conventional hematology gives results in absolute terms, whereas fluorescence cytometry results are in relative terms. Hematology analyzers deliver precise volumes and dilutions, and are thus able to measure absolute cell concentrations, or absolute counts of cell types per microliter of human blood. The fluorescence cytometry method gives only relative concentrations, or percentages of the various cell types.
A third distinction is that the hematology method is generally automated, whereas the fluorescence cytometric method as generally practiced today, is at best semi-automated, both in sample preparation, and in sample analysis. The fluorescence cytometry method is therefore significantly more labor intensive than the hematology method.
Both methods use cell by cell analysis. Therefore, due to the high concentration of cells in whole blood, it is necessary to dilute the blood samples prior to analysis so that individual cells can be isolated for sensing within a flowcell.
An example of an instrument for performing automated hematology measurements is the Cell-Dyn.RTM. 3000 instrument, which has been sold for several years by Sequoia-Turner, a predecessor in interest of Abbott Laboratories. The Cell-Dyn.RTM. 3000 instrument uses "impedance" measurements to count and size RBCs and PLTs, "absorption" measurements to determine the concentration of hemoglobin in RBCs (MCH), and "optical scatter" measurements to count and classify WBCs and the five part differential.
The Cell-Dyn.RTM. 3000 instrument automatically prepares blood samples, measures cell parameters and generates test results. The complete automation of sample preparation is such that no substantive operator intervention is required once the patient sample of whole blood has been presented to the analyzer. As mentioned previously, in order to assure accurate "patient counts" for the various cell classes, the Cell-Dyn.RTM. 3000 instrument provides precise sample volumes, reagent volumes and dilution volumes. Patient counts are generally defined as the number of "events" per microliter of blood. The events may be RBCs, PLTs, WBCs, and classes or subclasses thereof.
Other commercially available devices for performing hematology measurements include the Coulter.RTM. STKR, the Sysmex.RTM. NE8000, and the Technicon.RTM. H-1. Each of these uses combinations of scatter, impedance, and absorption to distinguish and quantify cells, and can thus be classified as a conventional hematology instrument.
In contrast, the fluorescence flow cytometer incorporates the principles of fluorescence cell analysis with light scatter. In general this requires that the cell be stained with an appropriate color dye, or that a fluorochrome label be attached to an antigen or antibody on the cell's surface thus indicating the occurrence of a specific antigen-antibody reaction.
In fluorescence flow cytometry, a suspension of previously stained or fluorescently labelled particles, typically cells in a blood or other biological fluid sample, is transported through a flowcell where the individual particles in the sample are illuminated with one or more focused light beams. One or more detectors detect the interaction between the light beam/s) and the labeled particles flowing through the flowcell. Commonly, some of the detectors are designed to measure fluorescent emissions, while other detectors measure scatter intensity or pulse duration. Thus, each particle that passes through the flowcell can be mapped into a feature space whose axes are the emission colors, light intensities, or other properties, i.e. scatter, measured by the detectors. Preferably, the different particles in the sample can be mapped into distinct and non-overlapping regions of the feature space, allowing each particle to be analyzed based on its mapping in the feature space. In this respect, flow cytometry differs from the conventional hematology instruments in that some of the feature space axis includes fluorescence emissions.
As noted above, lymphocyte subclasses are health determinants. Thus, it is desirable that these and other parameters be measured accurately. Although known hematology and fluorescent flow cytometry instruments have made significant advances in the ability to characterize blood cells, a problem still faced in this area is the difficulty in obtaining accurate patient count values for certain classes of cells.
An example of this problem is the CD4 cell patient count. Current analysis methods calculate the CD4 cell patient count from cell parameters measured on a hematology instrument and a separate fluorescence flow cytometry instrument. This calculation can provide up to 100% variability in absolute CD4 patient counts done on a single individual one week apart. See, e.g.: Update, Testing In The Blood Bank, Volume 5, No. 2, pages 1 to 6, published 1991 by Ortho Diagnostics Systems, Inc.
The following articles discuss additional difficulties with developing CD4 patient counts using current methods and devices;
The Lancet, Volume 340, Aug. 22, 1992, page 485 describes variation in CD4 count results when different analyzers are used. The variation appears to stem from different lymphocyte count results. PA1 Journal of Infectious Diseases, 1990, Volume 161, pages 356 to 357 describes variations in CD4 count due to variability in the reported lymphocyte concentration. The resulting variation in CD4 results has a deleterious effect on the patients' morale. PA1 Journal of Acquired Immune Deficiency Syndromes, 1990, Volume 3, No. 2, pages 144 to 181 reports large variations in CD4 counts for both HIV positive and control subjects. The fraction of lymphocytes that are CD4 positive is relatively constant, while the WBC count and the fraction of WBCs that are lymphocytes vary greatly. This variability points to the need for standardized analysis procedures. PA1 Laboratory Medicine, August 1983, Volume 14, No. 8, pages 509 to 514 discusses numerous spurious results and their causes in automated hematology analyzers.
The entire disclosure of each of the above-identified references is incorporated herein by reference.
One reason for variability in CD4 patient counts is manual sample preparation that cannot be controlled precisely and depends on operator proficiency. For example, a conventional procedure for determining a CD4 patient count starts with drawing two tubes of blood from a patient. One tube is analyzed on a hematology instrument which generates several measured and/or calculated parameters for the blood sample, including a total lymphocyte patient count, a lymphocyte percentage and a total WBC patient count. The second tube of blood is analyzed on a fluorescence flow cytometry instrument. The sample preparation steps for the flow cytometry tests are labor intensive and operator dependent. These steps do not readily lend themselves to automation and precision.
To prepare the sample for the flow cytometry instrument, the operator manually pipettes a volume of blood from the sample tube into an analysis tube. A volume of the desired fluorochrome labeled monoclonal antibody is added. The sample/antibody mixture is then incubated for a predetermined time at a predetermined temperature to allow antibody/antigen bindings to take place. After incubation, the operator adds a volume of RBC lyse to destroy the RBCs in the sample. Timing is important during the lysing stage. If the operator does not allow the lyse reaction to continue long enough, RBCs may remain in the sample and distort the measurements. If the operator allows the lyse reaction to continue for too long, the lyse may attack the WBCs.
After determining that the lyse reaction is complete, the operator centrifuges and washes the sample to remove any debris left over from lysed RBCs. The centrifuge/wash step may be performed several times until the operator is satisfied that the sample is sufficiently clean. Debris, red cell "stroma" can interfere with the detection processes of the typical flow cytometer. The sample now contains WBCs with antibodies bound to cells bearing the complementary surface antigens. The operator re-suspends the sample in a volume of fixative, and then passes the sample through the fluorescence flow cytometry instrument.
The fluorescence flow cytometry instrument generates only percentage values for lymphocyte subsets. This is at least partially due to the fact that the numerous manual dilutions and volume reductions performed during the sample preparation steps do not allow the isolation of a precise measurement volume. Thus, the fluorescence flow cytometry instrument identifies the CD4 positive helper-T cells as the percentage of lymphocytes which are both positive for CD3 (T cell marker), and positive for CD4(helper-T marker).
The CD4 patient count is then calculated using the following equations: EQU (%lymph/100).times.(WBC count)=lymph count (%helper-T in lymph/100).times.lymph count=CD4 count
The lymph count and the WBC patient count are taken from the hematology instrument, while the "% helper-T cells in lymph" value is taken from the fluorescence instrument after a correction factor is applied based on the flow cytometer mapping of scatter and fluorescence.
There are several problems with the current methods of calculating patient count values for lymphocyte subsets. First of all, the calculation is based on values obtained from separate instruments that each have their own calibration and overall separate functions. Additionally, different testing methods may be used on the different instruments.
Not only are hematology instrument measurements different from fluorescence instrument measurements, but also there may be variations in results obtained from different hematology instruments.
Previous attempts to automate sample preparation in fluorescence cytometry testing have only been partially successful. Such systems still require the operator to perform sample preparation steps such as separating lymphocytes from other peripheral blood cells by density gradient centrifugation, and/or lysing red cells, removing red cell ghosts and cell debris by centrifugation, or preserving the morphology of the remaining white cells by suspending the white cells in an isotonic saline solution containing appropriate fixatives. These operations generally require the operator to manually alter the volume of the sample, thus compromising sample volume precision which can be achieved with automated mechanical volume dispensers.
Another problem with the present technique of doing the measurements on separate instruments is that a relatively large volume of patient blood is needed to fill two tubes. This is a problem because of the increased likelihood that the blood will become hemolyzed (red cells destroyed) as larger amounts of blood are drawn. Additionally, it may not be advisable or possible to draw a sufficient amount of blood from certain patients.
In leukocyte analyses, it is desirable that all of the RBCs be lysed. Because RBCs outnumber WBCs by about 700 to 1, a small number of unlysed red cells may significantly distort white cell patient counts. Some reagents used to lyse red cells require too lengthy an incubation period to be practical in an automated clinical analyzer. For example, the Tris buffered ammonium chloride solution recommended by K. A. Murihead in Clinical Cytometry, Ann. N.Y. Acd. Sci., vol. 468, pp. 113-127 (1986) takes about 5 to 10 minutes to lyse red cells, which may be impractical for automation.
Furthermore, incomplete hemolysis with certain lyric reagents may result in red cell stroma that retain sufficient hemoglobin or particulate matter to generate high background patient counts in automated clinical electro-optical systems. When this occurs, it is usually necessary to remove the WBCs to be analyzed from the red cell stroma by centrifugation, a procedure that is a limiting factor when adapting a reagent system for automation.
Some currently used reagent systems require cytochemical staining of fixed WBCs before differential analysis. These systems require timed addition of multiple reagents and incubation periods and may not be generally adaptable for quantifying nucleated red cells or lymphocyte subsets. Furthermore, each step of reagent addition or other manipulation of a blood sample may decrease the precision of the final patient count obtained.
The earliest stage of RBC, the nucleated red cell, NRBC, when found in the peripheral blood on conventional hematology analyzers can be confused for a small lymphocyte, since the lysis will not destroy the nucleus of the NRBC. Because of the ratio of RBCs to WBCs, even a relatively small percentage of NRBCs can lead to substantial error in the WBC and lymphocyte count. This may be troublesome in neonate or pediatric samples, in which the presence of NRBCs in peripheral blood is a normal condition. For this reason, the laboratory may do manual slide inspections on some of these samples. Conventional hematology analyzers are only able to flag these samples by noting the spreading out of the usual lymphocyte scatter cluster. The manual inspection results in a count of the number of NRBCs per 100 nucleated cells. This percentage is then used to correct the analyzer WBC count as follows: EQU Corrected WBC count=Analyzer count(1-manual NRBC percentage/100)
Clearly the need exists for an accurate automated count of NRBCs.
Another important class of immature red blood cells are "reticulocytes" which typically contain detectable amounts of RNA. A manual method of identifying and counting reticulocytes involves precipitating the RNA with a stain. A smear is pulled from the stained blood and manually examined under a microscope. The precipitated RNA appears as intracellular dots or filaments. Reticulocyte % is determined by manually counting 1,000 RBCs under a microscope and dividing those qualifying as reticulocytes by 10. The reticulocyte patient count is derived from the RBC patient count according to the following equation: EQU Reticulocyte count=(RBC count).times.(percent reticulocytes)/100
Both the precision and the accuracy of this manual method are less than desirable. There may be considerable variation in identification of reticulocytes as well as variation in counting techniques. Accordingly, there is a need for a cell analysis system that addresses the deficiencies described above.